Dose-dependent ultrasound neuromodulation of the human brain
Benjamin R. Kop, Morteza Mohammadjavadi, Eva Feredoes, Martin T.W. Scott, Anthony M. Norcia, Elsa Fouragnan, Ryan T. Ash, Gary H. Glover, Kim Butts Pauly
2026 · 10.64898/2026.09.27.754841
Abstract
Transcranial ultrasound stimulation (TUS) can noninvasively modulate deep brain structures, but the relationship between delivered dose and neuromodulatory response remains poorly characterized in humans. Here, we show that ultrasound-induced tissue displacement measured in vivo predicts local and downstream functional neuromodulation. In 24 healthy adults, we combined magnetic resonance acoustic radiation force imaging (MR-ARFI) with concurrent TUS-fMRI in a double-blind, counterbalanced, within-subject design. Relative to active control stimulation, targeting the lateral geniculate nucleus (LGN) increased visually evoked BOLD responses in the LGN and ipsilateral primary visual cortex, localized to the retinotopic projection zone of the stimulated LGN site. Displacement predicted effect magnitude in both regions across participants and covaried with voxelwise responses within participants. Simulated in situ intensity showed no significant association with displacement or neuromodulatory response. These findings link in vivo mechanical displacement to functional modulation of a human thalamocortical circuit, providing an empirical basis for individualized dosing in causal neuroscience and clinical intervention.
Abstract via europepmc.
Exposures
Exposure 1: LGN-TUS (left lateral geniculate nucleus), with temporal grey matter as active control site
Target: lateral geniculate nucleus — “lateral geniculate nucleus of the thalamus (LGN)”
Device: IGT / Imasonic · Imasonic (transducer); Image Guided Therapy (generator) · 64-element phased array transducer; IGT generator ✓
| Waveform | continuous | |
|---|---|---|
| Fundamental frequency (kHz) | 500 | ✓✓✓ |
| Pulse duration (ms) | not applicable | ⚑ |
| Pulse repetition frequency (Hz) | not applicable | |
| Duty cycle (%) | not applicable | |
| Sonication duration (s) | 0.145 | ✓✓✓ |
| Free-field pressure (kPa) | not reported | |
|---|---|---|
| Free-field Isppa (W/cm²) | 50 | ✓✓✓⚑ |
| Free-field Ispta (W/cm²) | not reported | |
| In-situ estimate | not reported | |
| In-situ pressure (kPa) | not reported | |
| In-situ Isppa (W/cm²) | not reported | |
| In-situ Ispta (W/cm²) | not reported |
A 145 ms continuous pulse with 7 ms linear ramping, delivered six times per 15-second block time-locked to contrast decrements (870 ms per block); TUS in 6/7 of the epochs after an unstimulated first epoch, characterised by the authors as an average PRF of 0.46 Hz. Two 10-minute scans.
Flags from extraction
exposures[0].free_field.isppa_w_cm2— MR-ARFI scans used a higher free-water Isppa of 75 W/cm2 with 7 ms pulses; not recorded as an exposure.exposures[0].in_situ.method— Individualised k-Plan simulations were run but simulated in situ intensity values are not stated in the main text.randomised— Paper says counterbalanced order and pseudorandomized blocks; no explicit random allocation statement.exposures[0].timing.pulse_duration_ms— Single uninterrupted 145 ms burst per trial treated as continuous; the average PRF of 0.46 Hz is given only as a description.auditory_control— Masking white noise and a 4 kHz tone; blinding tested at end of session; the authors note the tone masked rather than matched the sound.sham_type— Control was active TUS to left temporal grey matter plus a no-TUS baseline condition.